论文标题:G.729A声码器算法研究和DSPs实现
Research on G.729A Vocoder Standard and Implementation on DSPs
论文作者 徐全元
论文导师 邵玉斌,论文学位 硕士,论文专业 通信与信息系统
论文单位 昆明理工大学,点击次数 103,论文页数 88页File Size4360k
2005-03-01论文网 http://www.lw23.com/lunwen_712383627/ G.729A;合成分析-线性预测编码;TMS320C5416 DSP芯片;汇编优化
G.729A; AbS-LPC; TMS320C5416 DSP; assembly optimization
本文主要目标是基于德州仪器公司(TI)的TMS320C5416 DSP芯片的ITU G. 729A 8kbit/s声码器实现。ITU提供了G. 729A声码器的16位定点ANSI C仿真程序。由于在编、解码过程中需要大量计算耗时,该ANSI C程序并不适合采用数字信号处理器(DSP)芯片来实时实现。在研究ITU G. 729A标准算法和学习并熟悉TMS320C5416 DSP开发后,将C代码转换为可实时实现的TMS320C5416汇编程序以及基于TMS320C5416-100 DSK硬件目标平台实现G. 729A声码器所需的硬件设计是本文的工作重点。因此,本文主要任务包括:研究ITU G. 729A标准算法和熟悉TMS320C5416 DSP开发;G. 729A声码器ANSI C代码仿真实现并转换为可实时实现的TMS320C5416汇编程序;TMS320C5416-100 DSK硬件目标平台的硬件程序设计。 G. 729A是ITU于1996年11月公布的G. 729建议的附件A。其中,ITU于1996年3月制定了G. 729语音编码方案,即采用共轭结构—代数码激励线性预测(CS-ACELP)技术的具有8kbit/s码速率的语音编码算法建议G. 729,该算法是以按合成分析法(AbS)和码激励线性预测(CELP)技术为基础提出的,属于ITU已制定语音压缩国际标准中最新且码率较低的一种。G. 729A主要目的是在基本不影响编码性能的情况下减少G. 729计算复杂度,它主要应用在个人移动通信、数字卫星系统和高质量数字移动无线通信等领域。 本文详细研究了G. 729A基于TMS320C5416 DSPs实时实现的设计与开发,在讨论了语音编码关键技术和G. 729A标准算法后,先对G. 729A实现的C程序进行分析,再重点论述了G. 729A在TMS320C5416型DSPs上实现的软、硬件设计开发的过程和技术。最后对该声码器进行性能分析。 DSP的软、硬件的设计和开发是G. 729A实现的核心技术。软件设计和开发过程中,本文强调的是C程序的移植和优化时的关键技术,如文件操作,汇编优化的方法,C和汇编函数互相调用的方法。在硬件设计和开发时主要考虑的问题是硬件中断编程以及McBSP和AD/DA技术。 本文共分为六章。第一章概述了语音数字压缩技术,第二、三章论述了语音编码的基本原理和G. 729A编解码器的算法,第四章研究了G. 729A的实现的C程序,第五
The primary object of this paper is to implement the ITU G.729A 8kbit/s vocoder based on TMS320C5416 DSPs. 16 bit fixed-point ANSI C code is available from ITU but it is not suitable for implementation in real time using a digital signal processor(DSP) due to the large computational time required for the encoder and decoder. The most vitalest part of this paper is to convert the ANSI C codes into TMS320C5416 assembly language and hardware design required for G.729A implementation on TMS320C5416-100 DSK hardware object plat . The main tasks of this paper include: understand G.729A standard; convert the ANSI C codes into TMS320C5416 assembly language; program for hardware of TMS320C5416-100 DSK hardware object plat.G.729A is ITU Recommendation G.729-Annex A and is publicized by ITU in Nov 1996.International Telecommunication Union(ITU) established a scheme of G.729 speech coding in March 1996. The ITU G.729 recommendation contains the description of an algorithm for coding of speech signals at 8kbit/s using Conjugate-Structure Algebraic-Code Excited Linear Prediction(CS-ACELP), which makes use of Analysis-by-Synthesis and Code Excited Linear Prediction in the scheme and belongs to the newest、 lower rate speech code of international standards established by ITU. The main aim of G.729A is to reduce computational complications. Mostly applied fields of G.729A include personal mobile communication, digital satellite system and high quality digital mobile wireless communication, etc.This paper detailedly studies the design and development of G.729A real time implementation based on TMS320C5416 DSPs. After discussing important techniques of speech coding and G.729A standard, firstly analyze ANSI C code of G.729A implementation, following to emphasis on discussing the software、 hardware design and development of G.729A real time implementation on TMS320C5416 DSP. At last, analyzing the performance of the vocoder.The software、 hardware design and development of DSP is core technique of G.729 A real time implementation. During software design and development, this paper emphasizeson transplant and optimization techniques of C code, such as file operation, assembly optimization, call of C and assembly function each other. The important questions are hardware interrupt programming, McBSP and AD/DA technique.This paper is organized as follows: Chapter 1 summarizes speech signal coding techniques. In Chapter 2 and 3, the speech coding rationales and G.729A encoder and decoder algorithms are discussed, respectively. Chapter 4 studies the software that defines this coder in 16 bit fixed-point arithmetic. Chapter 5 detailedly discusses the software. hardware design and development using TMS320C5416 DSP. This paper is summarized in the last Chapter 6.
|
